Evaluating Parameter Transfer in FALQON Across Graph Families
Abstract
We evaluate FALQON parameter transfer for Max-Cut, transferring sequences from small donors () to 14-node recipients. Using 3-regular and Erd\H{o}s-R\'enyi families, we show that transfer success is dictated by the recipient graph, not the donor. Transfer excels for dense recipients -- achieving high approximation ratios regardless of the donor -- but remains challenging in sparse cross-family cases. Crucially, performance is highly resilient to donor size, with 8-node donors matching larger instances. Thus, cheap small graphs can provide robust parameters for larger targets, significantly reducing the measurement overhead of the feedback loop.
Cite
@article{arxiv.2605.29917,
title = {Evaluating Parameter Transfer in FALQON Across Graph Families},
author = {Alisson dos Passos Fumaco and Marcos Vinicius Reballo and Fernando Augusto Caletti de Barros and Gabriel Fernandes Thomaz and Eduardo I. Duzzioni},
journal= {arXiv preprint arXiv:2605.29917},
year = {2026}
}
Comments
11 pages, 3 figures, Accepted at the SBCCQ - 1 Simp\'osio Brasileiro de Computa\c{c}\~ao e Comunica\c{c}\~ao Qu\^anticas